Looping correlation between variables

[Bibek Sapkota]

Hi,
With following data:

* Example generated by -dataex-. To install: ssc install dataex
clear
input float(disease_experience drought_experience flood_experience disease_prediction drought_prediction flood_prediction)
1 3 1 2 1 3
2 4 3 3 3 4
3 5 4 4 6 5
4 6 5 1 7 6
end

I executed

unab experience : *_experience
local experience: subinstr local experience "_experience" "", all

unab prediction : *prediction
local prediction : subinstr local prediction "_prediction" "", all

local prefixes : list experience & prediction

foreach s of local prefixes { 
pwcorr `s'_experience `s'_prediction, sig
}

Actually, I wanted a single matrix of all the coefficients and p values with *experience variables on row and *prediction variables on column. But, it produces as many matrices as the number of variables.

Would any tweaking in the code help me get what I wanted?

[Nick Cox]

ssc describe cpcorr
ssc install cpcorr
help cpcorr

P-values need more work. But you can look over the resulting matrix: the degrees of freedom are a constant for all cells.

[Bibek Sapkota]

Thanks. Much appreciated.

[Bibek Sapkota]

Hi,
Could we incorporate star (0.05) thing (i.e. to indicate if the correlation is significant) into the cpcorr syntax?
Thanks

[Nick Cox]

Who is we here?

I can only speak for myself as author of this command: I don't know how that could be done without complete rewriting of the command and I have no interest in doing it myself. Sorry if that is disappointing.

Feel free to clone the command under a different name with whatever acknowledgment fits the case.